Erno Borbely thinks so too: Linear Audio | your tech audio resource
"Although Erno Borbely must have designed a dozen or more phono preamps, he never tires of trying new avenues of circuit design to find that elusive holy grail. In this article he teamed up with Sigurd Ruschkowski to design a preamp that considers the phono cartridge as a current source, offering a very low impedance input to the cartridge. The article gives several options to use the basic amplifier in inverting and non-inverting mode and in single ended or balanced mode. The sensitive input stages all use (of course) JFET devices, with BJTs called in for the output stage and an opamp for the DC servo. The article concludes with sophisticated low noise shunt power supply regulator to power the preamp."
Jan
All this does is to make the L/R time constant be determined by the internal resistance only. It is often thought that this would be the way to get the largest tau, or somehow make the transducer become a current source, but as I pointed out a while back you can synthesize a negative resistance for loading that can mostly cancel the internal R, making tau larger still. Unfortunately this is difficult to do while preserving low nonlinear distortion.
Cordell explains how to load an MM cartridge with a smaller R than typical to partially realize the 75us RIAA pole pair, in his Linear Audio article on phono preamps.
I think this idea suffers a little from the expectation bias of the various authors.
Agree with you there Scott. Not to mention for the 75us tau realization, one needs to know the cartridge characteristics---or have the preamp determine it as part of a setup routine.
I have seen some say Oh this would be great (the zero input Z approach) as I won't have to worry about cartridge loading anymore!
The kind of loading, whether low Z, or high Z does not directly effect the S/N ratio. It can effect the bandwidth, however, with tape reproduce stages. A 'shorted' input is really better for the best tape reproduce, if the tape speed is high enough.
This is not a new technique, Ampex, for example, used it in 1963 with the MR-70 playback stage, one of the quietest tube circuits ever made. Studer used it in the A-80MK2, but did not use it in the MK3.(go figure)
The reason for the improvement with a 'shorted' type input is that it suppresses any tape head resonance, so more extended bandwidth is practical, without reducing S/N by using a lower output (inductance) head to extend the bandwidth.
This is not a new technique, Ampex, for example, used it in 1963 with the MR-70 playback stage, one of the quietest tube circuits ever made. Studer used it in the A-80MK2, but did not use it in the MK3.(go figure)
The reason for the improvement with a 'shorted' type input is that it suppresses any tape head resonance, so more extended bandwidth is practical, without reducing S/N by using a lower output (inductance) head to extend the bandwidth.
Last edited:
Thanks, John!
On the other hand, I still remember you saying that your first MC-prepre (JC-1?) didn't sound good enough when going into shorted input. Wouldn't the same apply to tape repro amp?
Regarding Studer - it was MkI, actually, that used this approach (500mH heads).
Apparently, current flowing thru the windings caused excessive crosstalk on multitrack machines, so beginning from MkII they've changed to 1:10 transformer, which was removed from the circuit completely by the time they got to MkIV.
Best,
Agree with you there Scott. Not to mention for the 75us tau realization, one needs to know the cartridge characteristics---or have the preamp determine it as part of a setup routine.
I have seen some say Oh this would be great (the zero input Z approach) as I won't have to worry about cartridge loading anymore!
I'm interested. what are the other problems with this approach for MM other than you have to obtain the cartridge parameters? The Cordell approach where you put a zero at 8kHz in the RIAA response and set cartridge loading to match seems on paper to offer a neat compromise.
Also if you have a test record/noise source to measure frequency response it does appear to be easier to adjust R than try and adjust R and C to get the resonance working for you rather than against you!
I guess I would ask: what are the advantages?
Bob does mention that he didn't go all the way to 2122 Hz because of noise issues with the smaller R required, which includes the attenuation incurred as well as the thermal noise of the resistor itself. There are ways around the latter too, but the question remains as to the advantages. One at least is that the overload margin at high frequencies can be enhanced, but there are other ways to do this as well.
When MC became popular for phono, one guy in Berkeley was experimenting with cartridge loading on his MC. He told me that he noticed the sound became clearer and more detailed with an ever lower load on his MC. I took this to mean lower distortion.... as you never get this listening result with ever higher distortion.
But, he complained the S/N would be worse, of course. So, I figured a MC pre-preamp design to actively load the cart with very low Z and have gain at the same time for low noise. That very low thd and noise circuit and was the first compl, push-pull ss configuration which was published a very long time ago (TAA).
Only recently did Borbeley also do a low input Z MC pre-pre but unfortunately he did not pursue the possible lowering of cartridge generated distortion in the over-all final result. I believe that cartridge loading/damping 'affect' was a contributor to the improved sound.... just like back in the Berkeley loading experiment days but without increased noise.
I think there is more to be learned in this direction... rather than ever fine tuning with the more familiar circuits.
THx-RNMarsh
Last edited:
I remember that stuff. Here's mine from 1977:
Attachments
John, if you refer to Bob's way to 'overload' the cartridge to automatically implement the first roll-off, there's another advantage which may be worth it: it straightens out the hf frequency response, getting rid of he usual response ripples from mech resonance which are normally used to extend the cart response.
Jan
Some MM carts don't take to it IME - well not to ultra low input impedance, sub 1R. Others thrive, with significant improvements in measured thd and certain types of noise. If an ideal zero ohm load is presented, there can be an awkward audioband L/R pole to work around, which is tidily addressed as set out here:
http://patentimages.storage.googleapis.com/pdfs/US4470020.pdf
That also sets out claims for mechanical damping benefits.
Another benefit for MM is total elimination of cable/preamp C loading as a contributing factor, and elimination of the LCR audioband resonance. IIRC this isn't mentioned in the filing above, but it is self-evident and verifiable IME.
Situation is simpler if L/R tau is above the audioband, as can readily be the case for MC. Thing with MC is limited access to expensive cartridges, so I have limited experience but the few cases I've trialled work well.
There's no question that any such audible effect is typically not expectation bias IME, for good reasons. Not least that effects are measurable, explainable, and for some cartridges audibly degenerate. Elimination of the LCR resonance in MM, for example, is readily audible for good reasons.
I will say that all of my research has been paper based as I don't have anything setup I can test, but if you look at New Lamps for Old you get a few plots for MMs and you can see that the Shure V15 was pretty nicely behaved into 47K and a total C of probably 450pF. But modern shures are nowhere near as flat and other modern MMs are all over the shop. The ortofon M2 is possibly the best behaved of the modern bunch but still has a 2dB peak at 10KHz. If you can move the resonant loading to a single variable (load R) you stand a much better chance of getting flat up to 20KHz.
I'm not convinced on the damping or the need for 50KHz frequency response, I'd just like to get the audio band within +/-1dB.
Best thing is if you have a preamp with the controversial von neuman 4th pole correction you have a handy zero to play with to try this out for free.
Not for the great unwashed, but for DIY it seems to be to have significant advantages in getting the frequency response flat. What have I missed?
I'm not convinced on the damping or the need for 50KHz frequency response, I'd just like to get the audio band within +/-1dB.
Best thing is if you have a preamp with the controversial von neuman 4th pole correction you have a handy zero to play with to try this out for free.
Not for the great unwashed, but for DIY it seems to be to have significant advantages in getting the frequency response flat. What have I missed?
Well, changes are not just restricted to frequency response by any means. In the case of MM, load current is increased by +40dB, and typically there are measurable thd and IM changes. The operating point for the generator can be significantly altered, for better or worse IME. Interface connection EM noise typically significantly improves. There is the prospect for mechanical damping interaction, though I have not confirmed and would weakly say anecdotal evidence is negative in the carts I have used.
Lastly, removing the LCR resonance does not necessarily leave a flat audioband f response - it depends on the cartridge. This is because there is always a mechanical (cantilever) self resonance, and sometimes this is devised to interact with the LCR resonance to extend the hf range before roll-off. Remove the LCR resonance, and the mechanical resonance is naked and observable. So it is important to work with carts which have mech resonance above the audioband if ultra flat f response is the goal. This approach can work beautifully, but f response is by no means the whole story.
To avoid mech resonance in the audioband, select carts with low effective tip mass - simply because (for better or worse) measurement of the resonant frequency is how effective tip mass seems to be derived. Whether this is mechanically 'correct' as an explanation of the resonance doesn't matter, it correlates well IME.
HTH!
- Status
- Not open for further replies.